csce 580 artificial intelligence ch.7: logical agents

UNIVERSITY OF SOUTH CAROLINA Department of Computer Science and Engineering

CSCE 580Artificial IntelligenceCh.7: Logical Agents

Fall 2008Marco Valtorta

[email protected]


Acknowledgment• The slides are based on the textbook [AIMA] and

other sources, including other fine textbooks and the accompanying slide sets

• The other textbooks I considered are:– David Poole, Alan Mackworth, and Randy Goebel.

Computational Intelligence: A Logical Approach. Oxford, 1998

• A second edition (by Poole and Mackworth) is under development. Dr. Poole allowed us to use a draft of it in this course

– Ivan Bratko. Prolog Programming for Artificial Intelligence, Third Edition. Addison-Wesley, 2001

• The fourth edition is under development– George F. Luger. Artificial Intelligence: Structures

and Strategies for Complex Problem Solving, Sixth Edition. Addison-Welsey, 2009


Parts of [AIMA]/[P]1. Artificial intelligence2. Problem-solving3. Knowledge and

reasoningCh. 7: Logical

Agents4. Planning5. Uncertain knowledge

and reasoning6. Learning7. Communication,

perceiving, and acting8. Conclusions

1. Agents in the World: What are agents and how can they be built?

2. Representing and Reasoning Ch. 5: Propositions

and Inference3. Learning and Planning4. Multi-agent Systems5. Reasoning about

Individuals and Relations

6. The Big Picture


Outline• Knowledge-based agents• Wumpus world• Logic in general - models and entailment• Propositional (Boolean) logic• Equivalence, validity, satisfiability• Inference rules and theorem proving

– forward chaining– backward chaining– resolution


Knowledge bases

• Knowledge base = set of sentences in a formal language• Declarative approach to building an agent (or other

system):– Tell it what it needs to know

• Then it can Ask itself what to do - answers should follow from the KB

• Agents can be viewed at the knowledge leveli.e., what they know, regardless of how implemented

• Or at the implementation level– i.e., data structures in KB and algorithms that

manipulate them


A simple knowledge-based agent

• The agent must be able to:– Represent states, actions, etc.– Incorporate new percepts– Update internal representations of the world– Deduce hidden properties of the world– Deduce appropriate actions


Wumpus World PEAS description

• Performance measure– gold +1000, death -1000– -1 per step, -10 for using the arrow

• Environment– Squares adjacent to wumpus are smelly– Squares adjacent to pit are breezy– Glitter iff gold is in the same square– Shooting kills wumpus if you are facing it– Shooting uses up the only arrow– Grabbing picks up gold if in same square– Releasing drops the gold in same square

• Sensors: Stench, Breeze, Glitter, Bump, Scream• Actuators: Left turn, Right turn, Forward, Grab, Release,

Shoot


Wumpus world characterization

• Fully Observable No – only local perception• Deterministic Yes – outcomes exactly

specified• Episodic No – sequential at the level of

actions• Static Yes – Wumpus and Pits do not move• Discrete Yes• Single-agent? Yes – Wumpus is essentially

a natural feature


Exploring a wumpus world


Logic in general• Logics are formal languages for representing

information such that conclusions can be drawn• Syntax defines the sentences in the language• Semantics define the "meaning" of sentences;

– i.e., define truth of a sentence in a world

• E.g., the language of arithmetic– x+2 ≥ y is a sentence; x2+y > {} is not a sentence– x+2 ≥ y is true iff the number x+2 is no less than

the number y– x+2 ≥ y is true in a world where x = 7, y = 1– x+2 ≥ y is false in a world where x = 0, y = 6


Entailment• Entailment means that one thing follows from

another:KB ╞ α

• Knowledge base KB entails sentence α if and only if α is true in all worlds where KB is true

– E.g., the KB containing “the Giants won” and “the Reds won” entails “Either the Giants won or the Reds won”

– E.g., x+y = 4 entails 4 = x+y– Entailment is a relationship between

sentences (i.e., syntax) that is based on semantics


Models• Logicians typically think in terms of models, which are

formally structured worlds with respect to which truth can be evaluated

• We say m is a model of a sentence α if α is true in m

• M(α) is the set of all models of α

• Then KB ╞ α iff M(KB) M(α)– E.g. KB = Giants won and Reds

won, α = Giants won


Entailment in the wumpus world

Situation after detecting nothing in [1,1], moving right, breeze in [2,1]

Consider possible models for KB assuming only pits

3 Boolean choices 8 possible models


Wumpus models


Wumpus models

• KB = wumpus-world rules + observations


Wumpus models

• KB = wumpus-world rules + observations• α1 = "[1,2] is safe", KB ╞ α1, proved by model checking


Wumpus models

• KB = wumpus-world rules + observations


Wumpus models

• KB = wumpus-world rules + observations• α2 = "[2,2] is safe", KB ╞ α2


Inference• KB ├i α = sentence α can be derived from KB by

procedure i• Soundness: i is sound if whenever KB ├i α, it is also true

that KB╞ α• Completeness: i is complete if whenever KB╞ α, it is also

true that KB ├i α • Preview: we will define a logic (first-order logic) which is

expressive enough to say almost anything of interest, and for which there exists a sound and complete inference procedure

• That is, the procedure will answer any question whose answer follows from what is known by the KB


Propositional logic: Syntax• Propositional logic is the simplest logic – illustrates

basic ideas

• The proposition symbols P1, P2 etc are sentences

– If S is a sentence, S is a sentence (negation)– If S1 and S2 are sentences, S1 S2 is a sentence

(conjunction)– If S1 and S2 are sentences, S1 S2 is a sentence

(disjunction)– If S1 and S2 are sentences, S1 S2 is a sentence

(implication)– If S1 and S2 are sentences, S1 S2 is a sentence

(biconditional)


Propositional logic: Semantics

Each interpretation (relational structure) specifies true/false for each proposition symbolE.g. P1,2 P2,2 P3,1 false true false

With these symbols, 8 possible models, can be enumerated automatically.Rules for evaluating truth with respect to a model m:

S is true iff S is false S1 S2 is true iff S1 is true and S2 is trueS1 S2 is true iff S1is true or S2 is trueS1 S2 is true iffS1 is false or S2 is true i.e., is false iff S1 is true and S2 is falseS1 S2 is true iffS1S2 is true andS2S1 is true

Simple recursive process evaluates an arbitrary sentence, e.g.,

P1,2 (P2,2 P3,1) = true (true false) = true true = true

The interpretations for which a formula is true are its models


Truth tables for connectives


Wumpus world sentencesLet Pi,j be true if there is a pit in [i, j].Let Bi,j be true if there is a breeze in [i, j].

P1,1B1,1B2,1

• "Pits cause breezes in adjacent squares"B1,1 (P1,2 P2,1)B2,1 (P1,1 P2,2 P3,1)

“A square is breezy if and only if there is an adjacent pit”


Truth tables for inference

See text (p.208)for R1…R5


Inference by enumeration• Depth-first enumeration of all models is sound and complete

• For n symbols, time complexity is O(2n), space complexity is O(n)


Logical equivalence• Two sentences are logically equivalent iff true in same

models: α ≡ ß iff α╞ β and β╞ α


Validity and satisfiabilityA sentence is valid if it is true in all models (i.e, it is a tautology),

e.g., True, A A, A A, (A (A B)) B

Validity is connected to inference via the Deduction Theorem:KB ╞ α if and only if (KB α) is valid

A sentence is satisfiable if it is true in some modele.g., A B, C

A sentence is unsatisfiable (i.e. it is a contradiction) if it is true in no modelse.g., AA

Satisfiability is connected to inference via the following:KB ╞ α if and only if (KB α) is unsatisfiable


Proof methods• Proof methods divide into (roughly) two kinds:

– Application of inference rules• Legitimate (sound) generation of new sentences from old• Proof = a sequence of inference rule applications

Can use inference rules as operators in a standard search algorithm

• Typically require transformation of sentences into a normal form

– Model checking• truth table enumeration (always exponential in n)• improved backtracking, e.g., Davis--Putnam-Logemann-

Loveland (DPLL)• heuristic search in model space (sound but incomplete)

e.g., min-conflicts-like hill-climbing algorithms


ResolutionConjunctive Normal Form (CNF)

conjunction of disjunctions of literalsclauses

E.g., (A B) (B C D)• Resolution inference rule (for CNF):

li … lk, m1 … mnli … li-1 li+1 … lk m1 … mj-1 mj+1 ... mn

where li and mj are complementary literals. E.g., P1,3 P2,2, P2,2

P1,3

• Resolution is sound and complete for propositional logic


ResolutionThe soundness of the resolution inference

rule can be shown by case analysis: If li is false, then

(li … li-1 li+1 … lk) is trueIf li is true, then

mj is false and thus (m1 … mj-1 mj+1 ... mn) is true

Overall, whatever the truth of li is, the resolvent

(li … li-1 li+1 … lk) (m1 … mj-1 mj+1 ... mn) is true


Conversion to CNFB1,1 (P1,2 P2,1)

1. Eliminate , replacing α β with (α β)(β α)(B1,1 (P1,2 P2,1)) ((P1,2 P2,1) B1,1)

2. Eliminate , replacing α β with α β.(B1,1 P1,2 P2,1) ((P1,2 P2,1) B1,1)

3. Move inwards using de Morgan's rules and double-negation:(B1,1 P1,2 P2,1) ((P1,2 P2,1) B1,1)

4. Apply distributivity law ( over ) and flatten:(B1,1 P1,2 P2,1) (P1,2 B1,1) (P2,1 B1,1)


Resolution algorithm• Proof by contradiction, i.e., show KBα unsatisfiable


Resolution example• KB = (B1,1 (P1,2 P2,1)) B1,1 α = P1,2


Horn Clauses• Horn clauses are clauses with at most one positive literal• Clauses with exactly one positive literal are called definite clauses

– Definite clauses with no negative literals are called facts– Definite clauses with one or more negative literals are called rules

• Clauses with no positive literals are called indefinite clauses or integrity constraints (in databases) or impossibility axioms (in model-based diagnosis) or goal clauses (in logic programming)

• Horn Form (restricted to definite clauses)KB = conjunction of definite Horn clauses

– Definite Horn clause = • proposition symbol; or• (conjunction of symbols) symbol

– E.g., C (B A) (C D B)• Modus Ponens (for Horn Form): complete for Horn KBs

α1, … ,αn, α1 … αn ββ

• Can be used with forward chaining or backward chaining.• These algorithms are very natural and run in linear time• They are well described in section 5.2 [P] (and associated slides)• The backward chaining algorithm is used by Prolog


Forward chaining• Idea: fire any rule whose premises are satisfied in the

KB,– add its conclusion to the KB, until query is found


Forward chaining algorithm

• Forward chaining is sound and complete for Horn KB


Forward chaining example


Proof of completeness• FC derives every atomic sentence that is

entailed by KB1. FC reaches a fixed point where no new

atomic sentences are derived2. Consider the final state as a model m,

assigning true/false to symbols3. Every clause in the original KB is true in m

a1 … ak b4. Hence m is a model of KB5. If KB╞ q, q is true in every model of KB,

including m


Backward chainingIdea: work backwards from the query q:

to prove q by BC,check if q is known already, orprove by BC all premises of some rule concluding q

Avoid loops: check if new subgoal is already on the goal stack

Avoid repeated work: check if new subgoal1. has already been proved true, or2. has already failed

Observation: this is the repeated application of a special kind of resolution, called unit resolution, in which one of the resolved clauses consists of a single literal; the query is a negative literal


Backward chaining example


Forward vs. backward chaining

• FC is data-driven, automatic, unconscious processing,– e.g., object recognition, routine decisions

• May do lots of work that is irrelevant to the goal

• BC is goal-driven, appropriate for problem-solving,– e.g., Where are my keys? How do I get into a PhD

program?

• Complexity of BC can be much less than linear in size of KB


Efficient propositional inference

Two families of efficient algorithms for propositional inference:

Complete backtracking search algorithms• DPLL algorithm (Davis, Putnam, Logemann, Loveland)• Incomplete local search algorithms

– WalkSAT algorithm


The DPLL algorithmDetermine if an input propositional logic sentence (in CNF) is

satisfiable.

Improvements over truth table enumeration:1. Early termination

A clause is true if any literal is true.A sentence is false if any clause is false.

2. Pure symbol heuristicPure symbol: always appears with the same "sign" in all clauses. e.g., In the three clauses (A B), (B C), (C A), A and B are

pure, C is impure. Make a pure symbol literal true.

3. Unit clause heuristicUnit clause: only one literal in the clauseThe only literal in a unit clause must be true.


The DPLL algorithm


The WalkSAT algorithm• Incomplete, local search algorithm• Evaluation function: The min-conflict heuristic of

minimizing the number of unsatisfied clauses• Balance between greediness and randomness


The WalkSAT algorithm


Hard satisfiability problems

• Consider random 3-CNF sentences. e.g.,(D B C) (B A C) (C B E) (E D B) (B E C)

m = number of clauses n = number of symbols

– Hard problems seem to cluster near m/n = 4.3 (critical point)



• Median runtime for 100 satisfiable random 3-CNF sentences, n = 50


Inference-based agents in the wumpus world

A wumpus-world agent using propositional logic:

P1,1 W1,1 Bx,y (Px,y+1 Px,y-1 Px+1,y Px-1,y) Sx,y (Wx,y+1 Wx,y-1 Wx+1,y Wx-1,y)W1,1 W1,2 … W4,4 W1,1 W1,2 W1,1 W1,3 …

64 distinct proposition symbols, 155 sentences


• KB contains "physics" sentences for every single square

• For every time t and every location [x,y],

Lx,y FacingRightt Forwardt Lx+1,y

• Rapid proliferation of clauses

Expressiveness limitation of propositional logic


Maintaining Location and Orientation

• KB contains “physics” sentences for every single square

• PL-Wumpus cheats – it keeps x,y & direction variables outside the KB.

• To keep them in the KB we would need propositional statement for every location. Also need to add time denotation to symbolsLt

x,y FacingRight t Forward t Ltx+1,y

FacingRight t TurnLeft t FacingRight t+1

• We need these statements in the initial KB for every location and for every time.

• This is tens of thousands of statements for time steps of [0,100]


• Reflex agent with State.• Formed of logical gates and registers (stores

a value)• Inputs are registers holding current percepts• Outputs are registers giving the action to

take• At each time step, inputs are set and signals

propagate through the circuit• Handles time ‘more satisfactorily’ than

previous agent. No need for a hundreds of rules encoding states

Circuit-based Agents


Example Circuit


Location Circuit

Need a similar circuit for each location register.


Unknown Information in Circuits

• Propositions Alive and Ltx+1,y are always known

• What about B1,2 ? Unknown at the beginning of Wumpus world simulation. This is OK in a propositional KB, but not OK in a circuit.

• Use two bits K(B1,2) and K(B1,2)• If both are false we know nothing! One of them is set by

visiting the square.K(B1,2)t K(B1,2)t-1 V (Lt

1,2 ^ Breeze t)• Pit in 4,4?

K( P4,4)t K( B3,4)t V K( B4,3)t

K( P4,4)t ( K(B3,4)t ^ K( P2,4)t ^ K( P3,3)t )V ( K(B4,3)t ^ K( P4,2)t ^ K( P3,3)t )

• Hairy Circuits, but still only a constant number of gates• Note: Assume that Pits cannot be close enough together such

that you can build a counter example to K(B1,2)t above


Example of 2-bit K(x) usage


Avoid cyclic circuits• So far all ‘feedback’ loops have a delay. Why? Otherwise

the circuit would go from being acyclic to cyclic• Physical cyclic circuits do not work and/or are unstable.

K(B4,4)t K(B4,4)t-1 V (Lt4,4 ^ Breeze t) V K(P3,4) t V K(P4,3) t

• K(P3,4) t and K(P4,3) t depend on breeziness in adjacent pits, and pits depend on more adjacent breeziness. The circuit would contain cycles

• These statements are not wrong, just not representable in a boolean circuit

• Thus the corrected acyclic (using direct observation) version is incomplete. The Circuit-based agent might know less than the corresponding inference based agent at that time

• Example: B1,1 => B2,2. This is OK for IBA, but not for CBA• A complete circuit can be built, but it would be much more

complex


IBA vs. CBA• Conciseness: Neither deals with time very well. Both are

very verbose in their own way. Adding more complex objects will swamp both types. Both are poorly suited to path-finding between safe squares (PL-Wumpus uses A* search to get around this)

• Computational Efficiency: Inference can take exponential time in the number of symbols. Evaluating a circuit is linear in size/depth of circuit. However in practice good inference algorithms are very quick

• Completeness: The incompleteness of CBA is deeper than acyclicity. For some environments a complete circuit must be exponentially larger than the IBA’s KB to execute in linear time. CBAs also forgets knowledge learned in previous times

• Ease: Both agents can require lots and lots of work to build. Many seemingly redundant statements or very large and ugly circuits.

• Hybrid???


Summary• Logical agents apply inference to a knowledge base to derive

new information and make decisions• Basic concepts of logic:

– syntax: formal structure of sentences– semantics: truth of sentences wrt models– entailment: necessary truth of one sentence given another– inference: deriving sentences from other sentences– soundness: derivations produce only entailed sentences– completeness: derivations can produce all entailed

sentences• Wumpus world requires the ability to represent partial and

negated information, reason by cases, etc.• Resolution is complete for propositional logic

Forward, backward chaining are linear-time, complete for Horn clauses

• Propositional logic lacks expressive power

csce 580 artificial intelligence ch.7: logical agents

Documents